A hypothesis explaining Alzheimer's disease, Parkinson's disease, and dementia with Lewy bodies overlap

Abstract

Lewy body‐involving diseases (LBD) are commonly associated with Parkinson's disease (PD) featuring voluntary movement inhibition, due to dopaminergic neuron dysfunction in the substantia nigra. PD is clinically tracked through Lewy bodies (LB), composed of insoluble α‐synuclein aggregates sequestered with organelles, particularly inside neurons. However, α‐synuclein pathology also appears in incidental LBD, Parkinson's disease dementia, and dementia with LB (DLB). Incomplete explanations address how these clinical pathologies interrelate, LBD etiology variability, and frequently overlapping α‐synuclein and Alzheimer's disease (AD) pathologies. We hypothesize that (1) chronic environmental insult exposure and (2) senescence(‐like) neuron accumulation contribute toward initiating and sustaining LBD; individual cell vulnerability determines either cell reactivity, death, or senescence in response to environmental insults. We predicate that parkinsonian and other neurodegenerative symptoms over LBD progression involve (3) co‐occurring AD pathologies, wherein dementia symptomology develops when synergistic glial senescence, tau hyperphosphorylation, and possible α‐synuclein aggregation reach into regions involved in AD progression.

Highlights

• Senescence burden is predicted to explain α‐synucleinopathy progression.

• Senescence and cell death are hypothesized to occur in α‐synucleinopathies.

• Sub‐apoptotic stress is proposed to induce senescence in α‐synucleinopathies.

• Neuronal senescence likely first spreads α‐synucleinopathies to new regions.

• Glial senescence likely underlies Parkinson's disease and Alzheimer's disease overlap.

1. INTRODUCTION

1.1. Introduction to Parkinson's disease and complex overlap with dementia

Parkinson's disease (PD), which occurs mostly as a late‐onset, environmental, or non‐familial form, involves motor deficits in tremors and rigidity. These symptoms mainly derive from dopaminergic (DA) neuron death in the midbrain's substantia nigra pars compacta (SNpc), and centrally involve alpha‐synuclein (α‐syn) proteinopathy. ¹ , ² , ³ Encoded by SNCA, α‐syn plays multiple roles in various organs, including modulating synaptic vesicle trafficking, dopamine release, and DNA repair. ⁴ , ⁵ α‐syn misfolds as a monomer, where serine 129 phosphorylation (pSer129) likely acts as an indicator of uncleared α‐syn aggregation into oligomers and subsequent fibrils. ⁶ , ⁷ Alongside promoting further α‐syn aggregation, α‐syn fibrils are often sequestered, particularly inside neurons, as a likely damage response. α‐syn fibrils are trapped with lysosomes, mitochondria, autophagosomes, lipids, lipofuscin granules, iron, and potentially neuromelanin to form intracellular Lewy neurites and Lewy bodies (LBs): ⁶ , ⁸ all post mortem hallmarks of PD pathology. ¹ , ² , ⁹ , ¹⁰ , ¹¹ While other neuronal types also carry LBs and die over Lewy body‐involving disease (LBD) progression, ² DA neuronal loss mainly causes motor deficits in LBD. Moreover, various genetic mutations or variants, including LRRK2 (involved in autophagy), SNCA (induces α‐syn overproduction), and GBA1 (encodes for lysosomal β‐glucocerebrosidase [GBA], leading to insufficient α‐syn degradation), contribute to PD risk. ¹²

PD progression and likely spread via α‐syn pathology has been traditionally tracked using Braak staging, which proposes that LBs first appear in the dorsal motor nucleus of the vagus nerve, glossopharyngeal nuclei, and the anterior olfactory nucleus. ¹³ Later propagation of α‐syn pathology then involves the medulla and pons, with subsequent stages thereafter involving the SNpc and finally the neocortex in advanced PD progression. Initial LB appearance in these areas reinforces that peripheral neurons are first affected. ¹ , ¹³ , ¹⁴ Numerous hypotheses making sense of Braak staging in PD have been posited, namely the “brain first,” “gut–brain” axis first, and “dual‐hit” hypotheses owing to co‐occurring olfactory and gut–brain axis routes. ¹³ , ¹⁴ Due to phosphorylated α‐syn and LB spread involvement within various brain and peripheral regions, olfaction deficits, disturbed sleep–wake cycles, and gastrointestinal and other autonomic dysfunction (i.e., representing non‐motor symptoms) also occur in patients with PD. ¹² , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸

PD features LB appearance and spread, but is better understood as belonging to a spectrum of LBD, including PD with dementia (PDD) and dementia with LBs (DLB). ¹ , ⁹ , ¹⁰ , ¹¹ , ¹⁹ Multiple system atrophy (MSA) will be excluded from our discussion, based on its different mechanisms underlying α‐syn pathology and heavier involvement of oligodendrocytes compared to LBD; while many of the arguments raised in our review likely also apply to MSA, other reviews better represent current and future research directions to understand the differing disease progression between LBD and MSA. ²⁰ , ²¹ , ²² PDD involves PD progression that advances into acquiring dementia or progressive cognitive decline. A majority of surviving patients with PD will develop dementia after a decade of exhibiting parkinsonian symptoms, ²³ where PD progression generally involves co‐occurring α‐syn pathology and hallmarks of Alzheimer's disease (AD) pathology in hyperphosphorylated tau (hp‐tau) aggregates. ²⁴ , ²⁵ , ²⁶ The biological definition of AD has been recently updated, ²⁷ wherein AD pathology is primarily characterized by amyloid beta (Aβ) and hp‐tau proteins, beginning as monomers and aggregating into more complex structures (Box 1). DLB also involves secondary α‐syn pathology and LB formation, with a primary progression of pathology and symptoms determined by AD pathology burden. ²⁴ , ²⁸ , ²⁹

Box 1. Braak staging in Alzheimer's disease and relation to Lewy body disease protein aggregates

While synaptic loss serves most strongly as a correlate for cognitive decline in Alzheimer's disease (AD), ³⁰ post mortem assessments of cognitive decline and neurodegeneration severity have been well correlated with a set of six Braak stages measuring hyperphosphorylated tau (hp‐tau) aggregation or “neurofibrillary pathology.” ³¹ , ³² Explicitly, Braak staging used to track AD progression (“AD Braak staging”) is not Braak staging used for categorizing Parkinson's disease (PD) progression. ¹³ Notably, early stages of AD progression preceding clinical symptomatic decline involve increased amyloid beta (Aβ) aggregation. ²⁷

In AD Braak stages I and II, neurofibrillary pathology begins and spreads from the transentorhinal cortex into the entorhinal cortex and hippocampus. ³¹ Stages III and IV involve increased lesions and pathologies in these temporal lobe regions, as well as novel spread into limbic structures, including the amygdala and insular cortex. Finally, stages V and VI affect the remaining superior temporal gyrus and neocortex. ³¹ Mild cognitive impairment has been established to be quite common for AD Braak stages III and IV, whereas stage IV+ is associated with an overt dementia diagnosis (Figure 1). ³¹ , ³² , ³³

Hp‐tau aggregation can occur in the form of insoluble hp‐tau fibrils, neuropil threads, and neurofibrillary tangles (NFTs). ³² These hp‐tau aggregates likely initially form within neurons, from upstream enzymes (including p38 mitogen‐activated protein kinases [p38 MAPK] and glycogen synthase kinase‐3 [GSK3]) phosphorylating tau at multiple sites and forming subsequent hp‐tau oligomers and more complex aggregates. ³² , ³⁴ , ³⁵ However, neuritic plaques, or amyloid plaques containing dystrophic neurites (implying synaptic loss), Aβ, and hp‐tau, have been shown causally in multiple mouse models and correlatively within AD patient brains to appear earlier in abundance within new regions versus NFT burden. ³² , ³⁴ , ³⁶ Although these points toward neuritic plaques as being a major driver of hp‐tau spread, ³⁶ diffuse amyloid plaques without hp‐tau burden do not correlate well with AD progression. ³²

To assemble amyloid plaques, regardless of diffuse or neuritic nature, Aβ aggregation must first occur from the oligomerization of misfolded Aβ monomers into insoluble fibrils. ³² , ³⁵ Interestingly, both hp‐tau and Aβ accelerate aggregation in acidic conditions; ³⁴ , ³⁷ both protein aggregates also likely promote the aggregation of the other. ³⁵ Physiologically, amyloid plaque formation likely results from microglial cells phagocytosing Aβ and condensing them into insoluble aggregates within an acidic, lysosomal environment. ³⁴ , ³⁸ Microglia then likely exocytose aggregates containing Aβ, hp‐tau, and dystrophic neurites to form denser amyloid plaques in the local environment, ³⁴ , ³⁸ thereby pointing toward dysfunctional microglial activity as a primary suspected driver in spreading hp‐tau aggregates advancing AD progression.

AD is predominantly known for Aβ and hp‐tau aggregates, but co‐occurring alpha‐synuclein (α‐syn) aggregates in Lewy bodies (LBs) have been found in AD patient brain samples even without the dementia with LB (DLB) diagnosis. ²⁸ Phosphorylated α‐syn particularly appear earliest, most prominently in the amygdala and hippocampus, in the form of deposits within dystrophic neurites found in neuritic plaques, and commonly in co‐existing LBs and NFTs (sometimes within the same neuron, as observed in the amygdala). ³⁹ , ⁴⁰ Henceforth, some subclassifications of AD include an “AD diagnosis with LB pathology” or “Lewy body variant of Alzheimer's.” ⁴¹ Notably, patients with AD can develop motor decline toward overt dementia progression. ⁴² , ⁴³

Relativeto PD  and PDD, patients with DLB possess significantly more Aβ deposition (including amyloid plaques and deposits into blood vessels) and hp‐tau pathology (including neuritic plaques and neurofibrillary tangles [NFTs]). ²⁹ , ⁴⁴ , ⁴⁵ , ⁴⁶ LBs can also include hp‐tau inclusions in patients with PD; ²⁶ , ⁴⁷ , ⁴⁸ patients with DLB relatively progress faster into death, ⁴⁹ , ⁵⁰ where DLB involves cognitive decline and dementia diagnosis preceding or appearing with parkinsonian symptoms. ⁵¹ Although molecular mechanisms have not been established, some neuronal populations are selectively vulnerable to developing both insoluble α‐syn and hp‐tau aggregates. ²⁶ , ⁴⁷ , ⁵² Patients with AD can also have significant, co‐occurring phosphorylated α‐syn pathology in multiple brain regions (Box 1). ⁵²

FIGURE 1.

Senescence spread can explain overlapping LBD and AD progression. Accumulating senescence is proposed to progress α‐syn fibril spread and inclusions containing α‐syn fibrils, such as LBs. Particularly, LBD is proposed to advance based more on neuronal senescence and connectome spread, whereas AD progresses based more on glial senescence spread. To differentiate PD and PDD, which have substantial α‐syn spread and parkinsonian motor symptoms preceding dementia symptoms, from DLB, PD and PDD proposedly follow the “body‐first” hypothesis under the SOC model. The SOC's “brain‐first” hypothesis postulates that LBs appear often foremost in the amygdala; this can be explained by senescent glial cell accumulation and co‐development of hp‐tau and α‐syn pathology, particularly in patients with dementia progression caused by AD. PDD and DLB diagnoses also likely follow a limbic predominant spread, due to the senescence burden and co‐occurring AD pathology accumulating earlier in Braak staging regions critical for dementia progression. In AD, senescence has been argued to advance dementia progression via local accumulation of oxidative stress, amyloid beta, and hp‐tau aggregates. As senescent cell clearance decreases over aging, inevitable overlap occurs between LBD and AD as each disease progresses. AD pathology likely follows a local regional Braak staging system of spread, where earlier overlap and exacerbated senescence buildup may also occur involving both AD and α‐syn pathologies. DLB likely involves pathology spread in both gut–brain route involvement and especially olfactory lobe; however, DLB is posited to favor a limbic predominant spread due to likely, increased involvement of AD pathology and affecting limbic regions involved in Braak staging. Faster progression of patients into death occurs in DLB versus patients with either (1) AD with minimal motor function decline, or (2) PD with slow or minimal cognitive decline. This is likely due to a predicted increased senescence burden in DLB, relative to other disease trajectories. Bidirectional arrows for PDD and DLB indicate that other areas may likely have accumulated significant senescence glial cell burden, due to co‐occurring AD pathology. Figure created with BioRender.com. α‐syn, α‐synuclein; AD, Alzheimer's disease; DA, dopaminergic; DLB, dementia with Lewy bodies; hp‐tau, hyperphosphorylated tau; ILBD, incidental Lewy body disease; LB, Lewy body; LBD, Lewy body‐involving disease; PD, Parkinson's disease; PDD, Parkinson's disease dementia; SOC, synuclein origin and connectome.

Regarding diagnostic assessment of α‐syn, hp‐tau, and Aβ biomarkers in patients, AD progression involves lowered cerebrospinal fluid (CSF) Aβ‐42 and greater hp‐tau relative to non‐demented controls. ⁵³ Here, plasma and CSF Aβ‐42/Aβ‐40 are lowered and correlate with reduced cognitive performance in DLB; ⁵⁴ , ⁵⁵ , ⁵⁶ CSF hp‐tau also likely predicts advanced stages of DLB and/or DLB subtypes with both cognitive impairment and high hp‐tau pathology. ⁴⁶ , ⁵⁴ , ⁵⁷ , ⁵⁸ Patients with PD or PDD additionally share a convergent decrease of CSF Aβ‐42, ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² but demonstrate inconsistent changes on CSF hp‐tau (i.e., increased or unchanged CSF hp‐tau vs. non‐demented individuals) relative to patients with AD. ²⁵ On the opposing end, patients with AD have demonstrated mixed findings on CSF α‐syn levels, ⁶³ but with a significant proportion of patients with AD demonstrating increased CSF α‐syn positively correlated with CSF hp‐tau. ⁶⁴ , ⁶⁵ , ⁶⁶ Seed amplification assays, which more sensitively measure α‐syn, have also consistently found that patients with AD containing enriched CSF α‐syn have worsened clinical prognoses and faster cognitive decline. ⁶⁷ , ⁶⁸ Altogether, during earlier disease progression and biomarker quantification aside from decreased CSF Aβ‐42, patients with PDD, DLB, or AD do not necessarily share in biomarker diagnosis; however, significant subsets of patients diagnosed with any single disease do overlap with pathologies from other LBD, and increased co‐presence of α‐syn, hp‐tau, and Aβ biomarkers leads to clinically worse outcomes. ⁶⁷ , ⁶⁸ Additionally, in advanced stages of disease progression, patients with AD or LBD likely develop mixed and overlapping pathologies. ²⁵

Furthermore, deceased individuals without neurodegenerative disease symptoms (whether somatic or mental) can display incidental LBD (ILBD). ⁹ , ¹⁰ Because not all patients with PD follow the proposed Braak staging, a more universal and accurate framework was proposed in the Unified Staging System for Lewy Body diseases (USSLBD). ⁹ , ¹⁰ , ¹⁴ The USSLBD uniquely accounts for ILBD and LB diseases by proposing that initial LB pathology is found in the olfactory bulb (USSLBD I), with subsequent spread affecting either brainstem structures in the medulla and pons (USSLBD IIa) or limbic areas including the amygdala and transentorhinal cortex (USSLBD IIb). USSLBD III features LB pathology in both brainstem and limbic regions, alongside a heavy involvement of the SNpc. Finally, the neocortex is involved at USSLBD IV. Post mortem samples from individuals with ILBD can have LB severity corresponding to any of USSLBD I through III, whereas only PD or DLB proceed to USSLBD IV. ⁹ , ¹⁰ This implies that ILBD overlaps with PD and DLB to a certain extent, where a “threshold” is required for parkinsonian symptoms.

A recent synuclein origin and connectome (SOC) model was also proposed to explain LBD, in which α‐syn spread would occur based on the importance of (1) the location of initial, significant α‐syn pathology, and (2) the synaptic connections that diseased neurons make with each other. ¹ The SOC enunciates “body‐first” and “brain‐first” hypotheses: the “body‐first” hypothesis predicts more bilateral spread of α‐syn aggregates, more olfactory deficits, and initial spread through the vagus nerve prior to involving the brain; the “brain‐first” hypothesis posits a relative, likelier slower spread and first location of significant α‐syn pathology in the brain (particularly the amygdala), with later spread to brainstem areas. ¹ , ¹⁴

Despite describing LBD and AD, it is less clear how these diseases work mechanistically, relate to each other, and integrate into a unified model explaining the heterogeneous clinical spread. Here, we provide a novel hypothesis that LBD and AD often overlap due to environmental insults and cellular senescence (hereby referred to in the rest of the text as “senescence”). Senescence is a functional state characterized by cell cycle arrest, nuclear lamina reorganization, and impaired organelle dysfunction enacted to evade apoptosis and increase apoptotic resistance. ⁶⁹ , ⁷⁰ These changes are often induced by transient p53 pathway activation and p16^INK4A as well as p21 upregulation, alongside more time‐permanent changes to the nuclear landscape, including via Lamin B1 downregulation and the formation of senescence‐associated heterochromatin foci (SAHF). ⁶⁸ , ⁷⁰ , ^{71 (p.21)} While neurons are postmitotic cells, different transcriptomic senescent signatures and signatures of multiple organelle dysfunction (including endolysosomal and mitochondrial pathways) distinguish in neurons typical postmitotic functioning to likely senescence across multiple neuronal types. ⁷² , ⁷³ Here, to add additional clarification, senescence differs from quiescence and postmitotic differentiation in cells; postmitotic cells can display a functionally senescent phenotype, and are not expected to display organelle dysfunction seen in senescent cells. ⁷² , ⁷⁴ , ⁷⁵

Furthermore, the central nervous system (CNS) also houses multiple glial types including the oligodendrocyte lineage cells (oligodendrocytes, oligodendrocyte progenitor cells), astrocytes, and microglia. Overall, glia altogether serve numerous key roles to support each other and neurons, ³⁴ , ⁷⁶ and can be generally divided into “states” including those with primarily “homeostatic” versus “maladaptive” functions. ⁷⁷ While we will discuss this topic below, the persistent effects of glia after becoming senescent likely also contribute to neurodegenerative diseases and dementia. ⁷⁸ With regard to AD and LBD, maladaptive microglia have been shown to seed singular proteinopathy aggregates (i.e., Aβ, hp‐tau, or α‐syn) via direct extracellular vesicle release, indirectly via promoting greater inflammatory signaling, and failing to clear and degrade preexisting pathologies; ³⁴ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ after mouse microglial cells were pretreated with Aβ , these Aβ‐treated microglial cells were transplanted into non‐transgenic mice and induced compound sporadic hp‐tau and α‐syn inclusions. ⁸⁷ Maladaptive astrocyte states also directly contribute to both AD and LBD, including via promoting inflammatory signaling and directly transmitting singular proteinopathies. ³⁴ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ , ⁹⁵ , ⁹⁶

1.2. LBD likely begins with enriched environmental insults, DA neuron selective vulnerability, and senescence

For DA neurons involved in LBD, LBs assemble intracellularly and remain intact only in surviving cells. ⁹⁷ Factors progressing LBD necessitate both neuronal death and separate neuronal survival mechanisms leading to LB formation. ⁹⁷ Certain neuronal types, especially DA neurons in the SNpc, may be selectively vulnerable (i.e., possessing traits rendering them susceptible to cell death or pathology, including via LB formation). ² In LBD, selective vulnerability in neurons includes increased axonal lengths, exacerbated intracellular calcium levels, and mitochondrial dysfunction, all traits that confer lower tolerance to oxidative stress. Although unhealthy when prolonged chronically, oxidative stress is a naturally occurring excess of reactive oxygen and nitrogen species (ROS and RONS) that overwhelm the cellular antioxidant systems, eliciting damage by altering proteins, lipids, and nucleotides (including DNA damage). ⁹⁸ RONS include hydrogen peroxide and peroxynitrite, and engage positive feedback loops increasing and increased by the secretion of cytokines including interleukin (IL)‐1, tumor necrosis factor alpha (TNF‐α), and interferon gamma (IFN‐γ). ⁶³ , ⁸⁵ Oxidative stress can further induce Aβ oligomer formation that accelerates α‐syn pathology. ³⁴ , ⁹⁹

In addition, oxidative stress independently from Aβ accelerates α‐syn phosphorylation and oligomerization toward aggregation ¹⁰⁰ , ¹⁰¹ that further depletes α‐syn levels. ⁵ , ⁷⁹ This aggregation also detracts from normative, healthy, and functional α‐syn roles in regulating DA synaptic vesicles and membrane binding. ¹⁰² Together, this putatively induces positive feedback loops of oxidative stress and neuronal death, where neurons undergoing apoptosis can release both α‐syn aggregates and neuromelanin (i.e., a dark pigment that is often synthesized from dopamine oxidation and is enriched in certain neuronal populations including DA neurons in the SNpc) inducing further local oxidative stress and inflammatory signaling. ⁸ , ¹⁰³ Most incidents of LBD including PD are not inherited, but instead result from environmental factors; ¹² α‐syn aggregate deposition has been observed first in either the olfactory bulb, ⁹ , ¹⁰ , ¹⁰⁴ or the gut–brain axis involving both the vagus nerve and enteric nervous system. ¹ , ¹⁰⁵ , ¹⁰⁶ Furthermore, these anatomical regions critically serve as first‐entry sites for environmental insults that can later affect main regions involved in PD (including the olfactory bulb and enteric nervous system). ¹⁰⁴ , ¹⁰⁷ Based on their ability to induce α‐syn pathology, DA neuron death, and parkinsonian symptoms, these environmental insults may include MPP (active neurotoxin derived from MPTP and used often in model organisms to approximate PD pathology), ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ paraquat, ¹¹¹ , ¹¹² , ¹¹³ rotenone, ¹⁰⁷ , ¹¹⁴ , ¹¹⁵ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸ heavy metals in lead and mercury, ¹¹⁹ , ¹²⁰ and potentially viruses including SARS‐CoV‐2 and herpes simplex virus ¹²¹ , ¹²² , ¹²³ , ¹²⁴ (Figure 2). ¹¹³ Finally, other gut‐involving diseases including inflammatory bowel disease and lifestyle factors may also increase the risk for PD. ¹²⁵ , ¹²⁶

FIGURE 2.

Environmental insults and neuronal involvement in Lewy body diseases. Putative effects of known risk factors for late‐onset LBD, most commonly in late‐onset PD, are overviewed in terms of outcomes on DA neurons. Environmental toxins, viruses, and heavy metals mostly kill DA neurons that are selectively vulnerable. This consequently releases monomeric and oligomeric α‐syn and if applicable, neuromelanin granules that amplify oxidative stress and inflammatory signaling in the local environment. In particular, these environmental insults may elicit increased oxidative stress inducing DNA double‐strand breaks and/or TNF‐α to damage neurons. While many neurons likely die, a smaller percentage of selectively vulnerable, DA neurons likely instead survive these insults, at the cost of (1) being depleted of functional α‐syn levels used for homeostatic neuronal functioning, and (2) forming α‐syn fibrils that coalesce into LBs. The act of LB formation is proposed to render the neuron alive, but accompanies (p53 pathway‐dependent) senescence induction and impaired organelle function. LBs also sequester lipofuscin granule accumulation and iron, two senescence markers. Thus, DA neurons containing LBs may be senescent. Finally, senescent DA neurons likely spread α‐syn aggregates to at least other neurons via a connectome approach. Even within a local brain environment, AD pathologies in Aβ and hp‐tau aggregates likely co‐occur and proposedly act as an environmental insult for LBD. Features resulting from environmental insults, including enriched oxidative stress and pro‐inflammatory cytokines in the local environment, can also induce further Aβ and hp‐tau pathology; both Aβ and hp‐tau are well known to synergize and promote further α‐syn aggregation and pathology in positive feedback loops, and hp‐tau can also induce senescence in surviving neurons. Across all figures, purple color is associated with agent senescence features; pink color indicates connotations with pathology and increased pro‐inflammatory signaling, whereas orange colors overview a more direct connection with oxidative stress. Question marks indicate mechanisms or suspected factors that require additional evidence for confirmation. Figure created with BioRender.com. α‐syn, α‐synuclein; Aβ, amyloid beta; AD, Alzheimer's disease; DA, dopaminergic; hp‐tau, hyperphosphorylated tau; LB, Lewy body; LBD, Lewy body‐involving disease; PD, Parkinson's disease; TNF‐α, tumor necrosis factor alpha.

Here, we posit that environmental insults trigger LBD progression through (1) increasing oxidative stress and (2) inducing cell death in selectively vulnerable, generally DA neuronal populations. ¹²⁷ Furthermore, these environmental insults and oxidative stress putatively result in α‐syn depletion within selectively vulnerable neurons. These neurons may attempt to survive via overexpressing α‐syn to buffer the oxidative stress, initially depleted α‐syn levels, and subsequent DNA damage elicited by environmental insults. ¹⁰³ , ¹²⁸ , ¹²⁹ Both GBA1 and SNCA contribute to increased PD risk; decreased levels of translated GBA protein lead to decreased clearance and increased overall α‐syn levels, ¹³⁰ whereas SNCA duplication has been shown to increase α‐syn levels. ¹² Albeit these findings, α‐syn overexpression has been demonstrated to induce downstream phosphorylation and subsequent aggregation into α‐syn fibrils. ¹³¹ Neurons have been shown to sequester α‐syn phosphorylation toward oligomerization and fibrils with organelles and other protein aggregates to form LBs, ⁶ , ¹³² which we posit represent a natural survival response to these environmental insults. ⁹⁷

While LB formation may be initially beneficial to prevent selectively vulnerable neuronal death, ⁹⁷ it allows for sequestering α‐syn fibrils and likely belies α‐syn aggregation. ⁶ This can lead toward α‐syn fibrils propagating and inducing α‐syn aggregation in other cells, including other neurons and oligodendrocytes, ⁸⁰ , ¹³² whereby continuing both processes concurrently results in surviving neurons that mediate local and retrograde α‐syn aggregate uptake. ² α‐syn spread is predicted to induce senescence in surviving and selectively vulnerable neurons, as a natural compensation for these neurons to avoid potential apoptosis (Figure 2). ¹¹⁸ As senescence characterization heavily relies upon cell type and environment, no universal markers for senescence exist. While senescence contributes to several natural and physiological roles (including wound healing and morphogenesis during development), ⁷⁰ it is relevant here in being generally characterized by phenotypic, consequent impairment. Senescent cells cannot break down endolysosomal products and develop mitochondrial dysfunction that increases RONS generation. ⁶⁹ , ¹³³

Although endocytosed and cell‐synthesized proteins vary per local environment and per cell type, resulting lysosomal swelling often leads to the accumulation of common substrates including senescence‐associated beta‐galactosidase (SA‐β‐gal), ferritin‐bound iron, and lipofuscin. ⁶⁹ , ⁷⁰ , ⁷¹ , ¹³⁴ , ¹³⁵ , ¹³⁶ , ¹³⁷ Senescent cells also commonly functionally adopt a senescence‐associated secretory phenotype (SASP) that turns neighbor cells senescent via paracrine secretion. This involves the appearance of cytoplasmic chromatin fragments, initiating the cyclic GMP‐AMP synthase, cyclic GMP‐AMP, and stimulator of interferon gene (cGAS‐cGAMP‐STING) pathway that initiates releasing type I interferons (IFN) and pro‐inflammatory cytokines, matrix metalloproteases (MMPs), high mobility group box protein 1 (HMGB1), and other secretions that vary across different cell types. ⁶⁹ , ¹³⁸ SASP regulation includes major transcription factors and proteins in nuclear factor kappa light chain enhancer of activated B cells (NF‐κβ), CCAAT‐enhancer‐binding protein (C/EBP), and HMGB1. ⁷⁰ , ¹³⁹

Senescence can be induced via multiple approaches: via replicative senescence in the Hayflick limit, DNA damage, oxidative stress via the p38 mitogen‐activated protein kinase (p38 MAPK) pathway, or the NLR family pyrin domain containing 3 (NLRP3) inflammasome. ⁶⁹ , ⁷⁰ As DNA damage and oxidative stress can also cause cell death, sufficient extreme severity of environmental insult exposure and individual cell resilience are proposed to induce either cell death or senescence (Figure 2). We predict that chronic, low‐dose exposure to environmental insults is likelier to induce senescence and resulting LBD etiology, which agrees with multiple PD pathology models. ¹¹² , ¹¹⁴ , ¹¹⁸ As we argue that senescence spread explains overlapping and unique patterns within LBD and AD, evidence here will focus on environmental insults inducing senescence.

In multiple rodent models and human models and samples (culture and post mortem tissue) representing DA(‐like) neurons, exposure to toxins, such as rotenone, and SARS‐CoV‐2 all each induced senescence via at least two confirmed markers of senescence: increased SA‐β‐gal staining (indicating endolysosomal dysfunction), higher p21 levels (indicating cell cycle arrest), and/or decreased Lamin B1 (which occurs early in senescence induction and cell cycle arrest). ¹⁴⁰ , ¹⁴¹ , ¹⁴² C/EBP, which is involved in inducing senescence, ⁶⁹ , ⁷⁰ was found to be required for α‐syn pathology spread in α‐syn overexpressing mice exposed to rotenone and upregulated in SNpc and gut biopsies collected from patients with PD. ¹¹⁸ Furthermore, each of these environmental insults has been linked to increased parkinsonian symptoms and progression into LBD, ¹² except for long‐term data assessing for COVID‐19. ¹²² Together, this implies that sublethal exposure to environmental insults can induce neuronal senescence. As neurons involved in LBD that die and spread α‐syn pathology are often selectively vulnerable to oxidative stress, ¹⁴³ we posit that these neuronal populations are more easily primed for entering either death or senescence due to exposure to oxidative stress induced by environmental insult exposure; such oxidative stress and subsequent cell death or senescence then further propagates α‐syn aggregation (Figure 2).

α‐yn phosphorylation and pathology were also shown to induce senescence, at least mechanistically via nuclear α‐syn accumulation and subsequent transcriptional alterations involving upregulating the p53–p21 pathway. ¹⁴⁴ Yoon et al. demonstrated that human SH‐SY5Y neuroblastoma cells overexpressing α‐syn increased SAHF and SA‐β‐gal positivity versus controls, where particularly p53–p21 pathway upregulation resulted in both SA‐β‐gal enrichment and α‐syn fibril accumulation in SH‐SY5Y cells. ¹²⁸ Proportionally induced α‐syn phosphorylation in these cells resulted in proportionally increased pATM and γH2A.X; as these markers indicate responses to DNA double‐strand breaks, this suggests α‐syn overexpression could induce senescence via DNA damage responses. Yoon et al. also observed senescence induction in a mutant A53T, human α‐syn expressing transgenic mouse model before the emergence of motor symptom deficits; they furthermore found that pSer129 α‐syn and γH2A.X levels were increased in the parietal cortex and hippocampi, indicating their effects have in vivo relevance. ¹²⁸ These findings were also confirmed by Shen et al. ¹⁴⁵ G2019S LRRK2, a frequent mutation associated with PD, ¹² , ¹⁴⁶ accompanied senescence via positive SA‐β‐gal staining and p21 induction in both SH‐SY5Y cells and G2019S LRRK2 transgenic mice. ¹⁴⁷ As both models involve α‐syn aggregate accumulation and transfer to other cells, this result further reinforces the hypothesis that senescence and neuronal α‐syn pathology are coupled.

SATB1 acts as a PD risk factor, where SATB1 knockout selectively induced several senescence markers in DA versus cortical neurons (all differentiated from human stem cells). ¹⁴⁸ In human stem cell cultures and in vivo midbrain DA mouse neurons, SATB1 downregulation induced p21‐dependent senescence. p21 upregulation was confirmed in the SNpc of patients with PD ¹⁴⁸ and SATB1–MIR22–GBA pathway disruption has been suggested to induce senescence in human stem cell–derived DA neurons. ¹³⁰ Altogether, environmental challenges may work to induce initial α‐syn fibril pathology, whereupon after α‐syn fibrils autonomously act as a separate mechanism that promotes subsequent neuronal senescence. As neurons spread α‐syn fibrils via a connectome model (i.e., synaptic networks connecting neurons), ¹ , ² this may act as a senescence spread mechanism involving neuron–neuron communication (Figure 2).

TNF‐α is elicited in response to α‐syn pathology, and also engages in positive feedback loops with oxidative stress. ⁷³ , ⁷⁹ , ¹⁴⁹ After collecting culture medium from primary mouse microglial cells exposed to pro‐inflammatory lipopolysaccharide, Bae et al. incubated SH‐SY5Y neuroblastoma cells and primary cortical mouse or rat neurons in the culture medium. ⁷³ In addition to eliciting the most α‐syn propagation among the tested cytokines, TNF‐α upregulated neuronal p53‐dependent senescence. Bae et al. also confirmed that these senescent neurons displayed a SASP profile and secreted increased α‐syn via lysosomal exocytosis, leading to an exacerbated pSer129 α‐syn appearance in nearby neurons. ⁷³ Environmental insults and oxidative stress may thus cause DNA damage and/or TNF‐inducing senescence and latent α‐syn aggregation in senescent neurons, ¹⁴⁷ entering a vicious feedback loop in which neurons surviving environmental insults are induced into feedback loops of senescence induction and further propagate α‐syn fibrils (Figure 2).

Abnormal iron exposure (via adding ammonium ferric citrate) and accumulation in induced pluripotent stem cell (iPSC)‐derived neurons were shown to trigger senescence via enriched SASP profiles and SA‐β‐gal staining. ⁷⁴ Said iron exposure was also shown to induce oxidative stress in these neurons, potentially linking another mechanism through which excess oxidative stress may relate to senescence induction in neurons. ¹⁵⁰ As LBs sequester iron, ⁸ LB neurons are likely senescent in ILBD, PD, and DLB. As ferric ammonium citrate additionally induces further α‐syn fibrillization and exacerbates senescence induction and oxidative stress, ¹⁴⁵ , ¹⁵¹ and intact neurons in PD also accumulate iron, senescent neurons with LB are predicted to similarly accumulate iron and become predisposed to propagating α‐syn pathology. Additionally, LB contains lipofuscin granules (organelles containing pigment and lipid‐containing residues left over from lysosomal digestion), ¹³ which serve as an additional senescence hallmark. ⁶⁹ , ¹³⁷ Therefore, the process of α‐syn aggregation itself likely sustains a senescence transition in neurons via multiple mechanisms for LBD progression.

Finally, α‐syn pathology was shown to co‐localize and synergistically increase hp‐tau aggregation and vice versa. ¹⁵² , ¹⁵³ Notably, different conformational strains of α‐syn fibrils have different capacities to seed hp‐tau inclusions; ¹⁵⁴ increased synergistic synuclein‐tau seeding may explain enriched LB and hp‐tau colocalization within a few brain regions, including the medulla oblongata and amygdala, in patients with AD or LBD. ²⁶ , ³⁹ , ⁴¹ , ⁴⁷ Both mice exposed to MPTP (to model PD pathology) and patients with PD also feature increased hp‐tau versus controls particularly in DA neurons of the striatum, hippocampus, and substantia nigra, ⁴⁸ , ¹⁰⁸ , ¹⁰⁹ with the appearance of insoluble hp‐tau aggregates being particularly featured in the MPTP model. Furthermore, risk factors and/or models for LBD all induce tau hyperphosphorylation beyond MPTP; this includes rotenone, paraquat, heavy metals, and viruses (potentially including SARS‐CoV‐2 and herpes simplex virus). ¹¹¹ , ¹¹⁵ , ¹²³ , ¹⁵⁵ , ¹⁵⁶ , ¹⁵⁷ , ¹⁵⁸ Considering that hp‐tau may inhibit apoptosis and lead to senescence, ³⁴ , ¹⁵⁹ we predict here that at least a subset of senescent neurons involved in LBD simultaneously develop α‐syn and hp‐tau pathology (Figure 2). Senescent neurons are additionally enriched in a P301L mouse model overexpressing human hp‐tau and in AD patient brains versus cognitively healthy controls; ⁷² , ¹⁶⁰ patient amygdala samples are notably enriched in neurons developing both LB and NFT aggregates in PD, AD, and DLB. ²⁴ , ²⁶ , ²⁸ , ⁴⁰ , ⁴¹ , ⁴⁷

Neurons with NFTs (both from the forebrain of transgenic P301L mice overexpressing human hp‐tau and unstated regions from post mortem brain samples of patients with progressive supranuclear palsy) are resistant to apoptosis; as these neurons with NFTs also retain intact their cell bodies over the disease progression of DLB and AD, and have upregulated Cdk2na implying increased p16 levels, it is proposed that hp‐tau pathology may induce neuronal senescence. ³⁴ , ¹⁶⁰ , ¹⁶¹ Phosphorylated α‐syn is known to induce tau hyperphosphorylation, ¹⁰⁸ , ¹⁵³ , ¹⁵⁴ and oxidative stress can induce p38 MAPK activation that leads to tau hyperphosphorylation. ³⁴ , ¹⁶² This may explain why a subset of LBs contain hp‐tau inclusions, ²⁴ , ²⁶ , ²⁸ , ⁴⁰ , ⁴¹ , ⁴⁷ as α‐syn fibrils sequestered within these LBs or certain neuronal populations may be primed for facilitating simultaneous hp‐tau seeding.

Furthermore, patients with PD, AD, or DLB all commonly demonstrate co‐occurring hp‐tau pathology and α‐syn aggregates in the brain. ²⁵ , ²⁸ , ⁴¹ , ⁴⁸ , ¹⁵³ Overall, hp‐tau pathways may provide an additional route toward inducing neuronal senescence in LBD and other dementias. We also predict here that the extent of comorbid hp‐tau pathology development depends on the selective vulnerability of the neurons exposed to the environmental insult, as codeveloping LB‐ and NFT‐containing neurons are enriched more in patients’ amygdala. ⁴⁰ Finally, as Aβ and hp‐tau can induce oxidative stress, DNA damage, and increase TNF‐α production independently from environmental insults (Figure 2), AD pathology itself is proposed to act as an environmental challenge that induces α‐syn pathology. ⁹⁹ , ¹⁵² , ¹⁵³ As environmental insults (including rotenone and paraquat exposure) can also directly induce Aβ and hp‐tau pathology, AD pathology likely also co‐occurs and simultaneously increases future AD pathology and α‐syn aggregation.

Overall, surviving vulnerable neurons in α‐synucleinopathies (PD, PDD, DLB) likely compensate for damage from environmental insults via upregulating α‐syn levels and becoming senescent to avoid apoptosis. ¹⁰³ , ¹⁶³ We posit that environmental insults (indirectly) promote α‐syn aggregate formation and further α‐syn spread and induce neuronal senescence and intracellular α‐syn aggregate formation relevant to LBD. Rotenone, paraquat, MPP+, and heavy metal models are notable for their ability to directly model PD pathology; ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹² , ¹¹⁴ , ¹¹⁵ , ¹¹⁶ , ¹¹⁷ , ¹¹⁹ , ¹⁵⁵ regardless of experimental model, we predict that each environmental insult has varying levels of epidemiological correlative support and likely can serve as additional, standardized in vitro/in vivo models of LBD. Together, environmental insults (including AD pathology) are proposed to cause the following alterations in selectively vulnerable neurons: increased death in some selectively vulnerable neurons, and increased survival in other such neurons leading into induced LB formation, depletion and aggregation of α‐syn (with possible co‐occurring hp‐tau), and forced α‐syn overexpression leading to neuronal senescence and continued, pathological α‐syn spread.

1.3. Environmental insults likely induce glial dysfunction and senescence in LBD

Environmental insults elicit TNF‐α and other cytokine release from pro‐inflammatory glial states. Although other glial cells and infiltrating immune cells contribute to α‐syn pathology, we will focus on astrocytes, microglia, and oligodendrocytes, considering their known involvement in LBD.

While explained elsewhere more comprehensively, these three glial cell types functionally converge into serving as non‐neuronal cell types that maintain a healthy CNS environment. ⁷⁶ This is mediated by complex glia–neuron and glia–glia interactions; ²² conversely, dysfunctional or suboptimal glial performance induces neuronal damage and death. ³⁴ Here, we predict that α‐syn is internalized and degraded by glial cells to minimize excessive α‐syn aggregation and downstream pathology. ¹⁶⁴ , ¹⁶⁵ Although microglia are known to be professional phagocytes generally with higher clearance capacities than astrocytes and the oligodendrocyte lineage cells, all glial cell types can demonstrate uptake of extracellular α‐syn monomers and aggregates from neurons. ²¹ , ⁸⁰ , ¹⁶⁴ , ¹⁶⁵ , ¹⁶⁶ Microglia and astrocytes containing α‐syn inclusions were also observed in post mortem olfactory bulb samples of patients with PD. ¹⁶⁷ α‐syn inclusions were also found in oligodendrocytes in a mouse model injected with preformed α‐syn fibrils, although with a relatively delayed accumulation versus neuronal α‐syn accumulation; ¹⁶⁸ this implies that α‐syn aggregation and propagation could occur faster in neurons relative to at least oligodendrocytes. Internalizing α‐syn monomers, oligomers, and fibrils further elicit inflammatory signaling in astrocytes and microglia, ⁸¹ , ⁸² , ⁸⁸ , ⁸⁹ where both glial types release TNF‐α. ⁷⁹ , ⁸⁹  TNF‐α acts potently toward inducing neuronal death or senescence in selectively vulnerable cells. ⁷³ Furthermore, pro‐inflammatory cytokines were found to independently upregulate α‐syn levels and encourage further aggregation. ⁷⁹

In mice overexpressing mutant SNCA G420A (leading to increased pSer129 α‐syn) exclusively in microglia and macrophages, et al. found resulting DA neuron loss. ⁸¹ The authors concluded that mutant SNCA G420A microglia displayed a hypophagocytic state (i.e., a reduced ability to phagocytose) after exposure to either IFN‐γ, exogenous α‐syn fibrils, or expressing pSer129 α‐syn. Finally, the authors found that these immune cells exposed to one or more challenges displayed excessive RONS release; exposure to both IFN‐γ and α‐syn pathology led to synergistically higher release of nitric oxide and increased oxidative stress. ⁸¹ The medium from primary mouse microglial cells, either exposed to α‐syn fibrils or IFN‐γ, was also extracted and revealed to induce oxidative stress in these cells when co‐cultured with DA neurons. The same medium caused neuronal death when the microglial cells were simultaneously exposed to IFN‐γ and α‐syn fibrils. ⁸¹ Furthermore, the microglial cells were found to release TNF‐α which selectively exacerbated both α‐syn pathology propagation and senescence in primary rodent neurons and SH‐SY5Y neuroblastoma cells. ⁷³

While this finding demonstrates the distinct effects of microglia in eliciting reactive states and resulting neuronal pathology relevant to LBD, it demonstrates how positive feedback loops of pro‐inflammatory cytokines, including IFN‐γ, and α‐syn pathology can be further mediated by glial cells. While this likely occurs in other cell types, pro‐inflammatory effects and positive feedback loops encouraging further α‐syn pathology were particularly documented in astrocytes. ⁸⁹ It is also well established that AD pathology involves a buildup of oxidative stress, hp‐tau, and Aβ aggregates via pro‐inflammatory cytokine exposure. ³⁴ As there is no reason that environmental insults and glial reactions happen exclusively in PD or AD, it sustains that in a complex biological system, pathologies initially thought as relevant to PD (α‐syn aggregates) should also occur in AD and vice versa. Particularly, DLB likely involves accelerated progression due to a likely enriched combination of pro‐inflammatory cytokine exposure and reactive glial states resulting in higher levels of oxidative stress, α‐syn, hp‐tau, and Aβ pathology versus either AD or PD.

As TNF‐α induces DA neuron senescence, ⁷³ and pro‐inflammatory glial states can release cytokines (including TNF‐α and IFN‐γ), whether directly or indirectly, in response to environmental insults, glial dysfunction also likely advances LBD progression via increased total neuronal counts of death and senescence. Other cytokines including IFN‐γ and IL‐1β are released by both glial cell types in response to α‐syn, ⁷³ , ⁸⁹ creating an inflammatory environment for neuronal and non‐neuronal cells (including cells forming the neurovascular unit, etc.) alike. As pro‐inflammatory cytokine release by one glial cell type can elicit further glial reactivity and release of other pro‐inflammatory cytokines, likely subsequent positive reinforcement loops of inflammatory signaling act as another central mechanism by which LBD progression occurs (Figure 3).

FIGURE 3.

α‐Synuclein transfer and likely glial senescence in Lewy body diseases. Environmental insult effects in LBD and α‐syn are explored in glial cells. Alongside killing and inducing likely senescence features in DA neurons, environmental insults specific for LBD may cause general pro‐inflammatory cytokine release and oxidative stress inducing inflammatory glial states and glial senescence. Both astrocyte senescence and pro‐inflammatory glial states may further damage the local environment, and are proposed to induce “priming” of non‐senescent microglia and astrocytes. These proposed primed glial cells, especially microglia, may prematurely engulf and kill senescent DA neurons containing at least insoluble α‐syn aggregates. As oxidative stress also induces simultaneous hp‐tau pathology and susceptibility of neurons to being prematurely phagocytosed, neurons with senescent features may also contain α‐syn aggregates co‐seeded with or containing independently seeded hp‐tau aggregates. Finally, α‐syn fibrils may act as another potential, unelucidated mechanism toward inducing senescence in astrocytes and microglia. If primed astrocytes and especially microglia prematurely phagocytose susceptible, senescent neurons containing both α‐syn fibrils and hp‐tau, this process, called phagoptosis, may act as one mechanism toward uptake of these insoluble aggregates and inducing senescence in microglia and astrocytes. Finally, senescence accompanies endolysosomal and mitochondrial impairment. Thus, phagoptosis and possible exocytosis of α‐syn and/or hp‐tau aggregates by senescent glia and neurons is proposed to further indirectly and directly promote α‐syn fibril and hp‐tau seeding, formation, and uptake by other local cells such as oligodendrocytes. In LBD, these proposed mechanisms are predicted to induce a vicious feedback loop of paracrine glial senescence, α‐syn aggregate spreading, and varying extents of co‐morbid hp‐tau pathology depending on the particular susceptibility of the DA neurons exposed to environmental insults. Figure created with BioRender.com. α‐syn, α‐synuclein; DA, dopaminergic; hp‐tau, hyperphosphorylated tau; LBD, Lewy body‐involving disease.

The exposure to extreme environmental insults (including paraquat and rotenone) can induce apoptosis in astrocytes, oligodendrocytes, and microglia, potentially via increased oxidative stress and pyroptosis (programmed cell death via inflammasome activation). ¹¹² , ¹⁶⁹ , ¹⁷⁰ As all three glial cell types internalize α‐syn, ²¹ , ⁸⁰ , ¹⁶⁴ , ¹⁶⁵ , ¹⁶⁶ glial death resulting from environmental insults putatively decreases the overall homeostatic support and contributes to oxidative stress and α‐syn release into the local environment. Furthermore, neuromelanin from human SNpc normally sequesters iron, lipofuscin, and other undegradable products to protect SNpc DA neurons by minimizing oxidative stress. ¹⁷¹ , ¹⁷² However, when neuronal death is initiated by environmental insults (Figure 2), extracellular neuromelanin is released and incites microglial release of TNF‐α and IFN‐γ. ¹⁷³ Therefore, exacerbated inflammatory signaling and pro‐inflammatory cytokine exposure, particularly TNF‐α released by glial cells, likely induce further oxidative stress, increased counts of separate cell death and senescence in selectively vulnerable neurons, and α‐syn pathology (Figure 3).

We predict that while such damage is first minimized by supportive and anti‐inflammatory cytokine release by homeostatic glial states, ¹⁶⁴ , ¹⁶⁵ time accrued over aging likely leads to repeated, chronic exposure to environmental and physiological insults, resulting in chronic inflammatory signaling (Figure 4). While the effects of aging on glial function are extensively reviewed elsewhere, ³⁴ this generally leads to an increase of: (1) pro‐inflammatory glial states, (2) declining homeostatic functions, and (3) glial “priming” and phagocytosis of neurons exposed to oxidative stress (notably via translocated phosphatidylserine to the outer cell membrane layer). ³⁴ , ¹⁷⁴ Furthermore, chronic low‐dose exposure to environmental insults likely captures LBD pathology and progression better over time; ¹¹² , ¹¹⁴ environmental insults likely also add to biological aging by accumulating additional oxidative stress, and increased separate cell counts of either induced death or senescence. Emergent aging hallmarks (including oxidative stress, pro‐inflammatory cytokine exposure and resulting inflammatory signaling, and increased senescence) also generate increased Aβ aggregate burden, ¹⁷⁵ which are all likely to accelerate additionally α‐syn buildup and pathology. ⁹⁹ Furthermore, we suspect that aging acts as a contributing threshold for declining homeostatic glial function in containing α‐syn pathology and co‐occurring AD aggregates in LBD. ¹¹⁰ , ¹¹⁶ , ¹⁷⁶ , ¹⁷⁷ , ¹⁷⁸ , ¹⁷⁹ , ¹⁸⁰ , ¹⁸¹ We propose that this decline, particularly resulting in relative accumulating neuronal and glial senescence, is what allows LBD to locally and heterogeneously progress per individual into end‐stage PD, PDD, DLB, and/or AD (Figure 1).

FIGURE 4.

Aging and senescence spread in explaining environmental PD. The hypothesis's postulates are applied to better account for PD progression. (1) Senescence by α‐syn fibril and/or hp‐tau aggregates, and (2) local oxidative stress and inflammatory signaling serve as central participants facilitating progression of PD; this can occur via a main connectome (neuronal) approach and/or an initiating/supporting paracrine (glial) spread to build up further senescence. Over healthy adulthood, these factors are putatively minimized; selectively vulnerable DA neurons are supported by homeostatic glial states tolerating α‐syn pathology and ILBD, and senescent cells are cleared and phagocytosed by immune cells. However, over aging and exposure to numerous environmental insults, homeostatic glial states likely become inadequate in managing increasing α‐syn and Alzheimer's disease pathology formation. This leads into beginning PD progression. Models of PD spread, under the SOC model and USSLBD, suggests that PD progression initially begins as asymptomatic ILBD from an origin source. Based on the SOC's “body‐first” hypothesis (i.e., α‐syn pathology beginning outside the brain) and ILBD, the hypothesis predicts that α‐syn pathology can begin distinctly in either the gut or olfactory bulb; as environmental insults and senescence spread require connected environments, while early, independent progression of both routes can occur simultaneously. Both routes eventually involve brainstem areas (e.g., raphe nuclei, locus coeruleus, and substantia nigra). The gut–brain route likely propagates α‐syn fibrils via a brainstem‐predominant spread (USSLBD IIa) first, involving neurons in the dorsal motor nucleus of the vagus nerve; subsequent LB progression in PD would involve limbic regions, including the amygdala and transentorhinal cortex, and neocortical spread in the frontal, temporal, and parietal lobes. α‐syn aggregate spreading and senescence buildup are also predicted to involve regions critical to dementia prognosis and progression (Figure 1); patients with olfactory route, “body‐first” PD are predicted to have increased dementia risk. Figure created with BioRender.com. α‐syn, α‐synuclein; DA, dopaminergic; hp‐tau, hyperphosphorylated tau; ILBD, incidental Lewy body disease; LB, Lewy body; LBD, Lewy body‐involving diseases; PD, Parkinson's disease; SOC, synuclein origin and connectome; USSLBD, Unified Staging System for Lewy Body Disorders.

Even a small percentage of senescent glia can exert disproportionately damaging and non‐homeostatic influences on other cell populations. ³⁴ On top of senescent cells inducing paracrine senescence in nearby cells, ⁶⁹ , ⁷⁰ senescence can lead to multiple pro‐inflammatory and anti‐supportive effects, including a decreased clearance and an increased spreading of pathological aggregates. ³⁴ , ¹³⁸ Senescent astrocytes release SASP factors and contribute to cognitive impairment via releasing HMGB1 and activating the NLRP3 inflammasome, ¹⁸² which can ultimately promote pro‐inflammatory microglial states and decrease oligodendrocyte precursor cell differentiation leading to lowered axonal myelination (which is affected in LBD and AD). ³⁴ , ¹⁷⁵ , ¹⁸³ Senescent astrocytes also directly downregulate EAAT1/2 transporters and promote neuronal death via extracellular glutamate and excitotoxicity; ¹⁸⁴ as excessive glutamate can induce oxidative stress and subsequent DA neuron death, ¹⁸⁵ senescent astrocytes may contribute to LBD progression.

In a MPTP mouse model, both primary cultured astrocytic cells and astrocytes in the SNpc displayed greater SA‐β‐gal and p16^INK4a immunostaining, decreased Lamin B1, and a SASP profile. ⁹⁰ Xia et al. pretreated both primary mouse astrocytic cells and the SNpc of MPTP+‐treated mice with astragaloside‐IV (a drug used in traditional Chinese medicine commonly for infections); this subsequently prevented MPP+‐induced senescence in both models, and rescued both DA neuronal populations and MPTP‐induced motor deficits. Here, the authors demonstrated that astragaloside‐IV worked at least through promoting autophagy and decreasing dysfunctional mitochondrial‐derived RONS. ⁹⁰ Although astragaloside‐IV gathered evidence for preventing senescence (instead of acting solely as a senolytic drug to selectively kill senescent cells), this causally reinforces that glial reactivity and senescence contribute to motor neurodegeneration seen in LBD. Autophagy deficiency (via knocking out ATG7) in CX3CR1+ cells (mainly microglia in the brain) was also found to prevent microglial senescence in the 5xFAD mouse model overexpressing Aβ. ⁷²

In contrast to high doses of paraquat inducing human iPSC‐derived astrocyte death, chronic low dosages of paraquat induced human iPSC‐derived astrocyte senescence via increased SA‐β‐gal and p16^INK4a immunostaining, and increased IL‐6 release as measured by enzyme‐linked immunosorbent assay. ¹¹² Chinta et al. confirmed that conditioned medium from these senescent astrocytes decreased DA neuron culture viability and proliferative capacities of neural stem cells, reinforcing that senescent glia can disproportionately harm healthy function of other cell populations. When p16^INK4a‐overexpressing cells (enriched in senescence, including senescent astrocytes) were selectively ablated in mice, this prevented decreased neurogenesis in paraquat‐exposed mice versus controls lacking senescent cell ablation; this led the authors to conclude that astrocyte senescence contributes to LBD pathology and neurodegenerative motor symptoms. ¹¹² After treatment with H₂O₂ (i.e., inducing oxidative stress) or rotenone, human fetal astrocytes displayed several senescence markers including increased SA‐β‐gal staining, p21 levels, multiple transcriptomic SASP markers, and gene enrichment conferring pro‐inflammatory roles. ¹⁸⁶ Conditioned medium from these human senescent astrocytes also exacerbated cell death in selectively vulnerable DA neurons (with SNCA duplication) exposed to rotenone.

Finally, senescent mouse astrocytes were shown to contribute to neurodegeneration via cGAS‐cGAMP‐STING signaling axis delayed PD‐like pathology in MPTP‐treated mice. ¹⁸⁷ As PD risk factors often involve chronic exposure to environmental insults, these insults in LBD may lead to astrocyte senescence contributing toward pathology in selectively vulnerable neuronal populations, which results in subsequent α‐syn propagation. Furthermore, on top of (1) oxidative stress, astrocyte senescence can be induced by overexposure to (2) hp‐tau and (3) Aβ aggregates. ³⁴ , ¹⁸² Provided earlier that environmental insults can (eventually) induce all three factors, one indirect mechanism by which other environmental insults (Figure 2) may induce astrocyte senescence could involve the buildup of AD pathology through increased environmental insult exposure.

Both MPP+ and α‐syn fibrils increased p21 in primary mouse astrocytic and microglial cell cultures (suggesting induced astrocyte and microglial senescence). ¹⁴⁰ Injection of α‐syn fibrils into both non‐transgenic mouse neocortex and SNpc also led to local increased immunostaining intensity of p21, and decreased intensities of Lamin B1 and HMGB1 in both astrocytes and microglia; ¹⁴⁰ this confirms multiple senescence markers in vivo for both glial types, induced (in)directly by α‐syn pathology. When mouse microglia were specifically induced to produce extra α‐syn via SNCA upregulation and targeting CX3CR1 promoter, these microglia decreased morphological complexity and displayed hypofunctionality in phagocytosing further α‐syn. ⁸¹ As senescent mouse microglia have displayed via hypofunctionality in phagocytosing extracellular hp‐tau aggregates, ⁸³ exhausted microglia failing to clear further α‐syn may also likely be senescent. Finally, primary astrocyte and microglial cell cultures derived from mouse cortex were made senescent via replicative division (displaying increased senescence markers of p16^INK4a and SA‐β‐gal enrichment, and downregulated Lamin B1), and were found to have dysfunctional autophagy and impaired α‐syn aggregate clearance. ¹⁸⁸

Senescent microglia likely contribute to sustaining harmful environments that disrupt supportive glial cell functions, ³⁴ while simultaneously causing α‐syn pathology in LBD. ⁷⁹ In A53T (expressing mutant α‐syn with a high aggregation rate relative to non‐transgenic α‐syn) mice, inducing p16^INK4a‐immunopositive microglia with light exposure and chiral nanogold particles significantly improved rotarod and wheel running performance; the authors support that removing senescent microglia decreased motor neurodegeneration after LBD progression. ¹⁸⁹ Finally, microglia overloaded with iron take on a dystrophic morphology. Dystrophic microglia, which we have defined as a phenotype with (1) ferritin upregulation and (2) progressively reduced morphological process complexity (later advancing into spheroidal swellings and possible cytoplasmic deterioration at least in humans), are arguably senescent (Box 2). ³⁴ , ¹⁹⁰ , ¹⁹¹

Box 2. Hypothesizing Alzheimer's disease as an aging disease of glial senescence

Briefly, we have argued that Alzheimer's disease (AD) is an aging disease that progresses via glial (and particularly microglial) senescence accumulation and spreading within key brain regions involved in AD Braak staging ³⁴ (Figure 1). Regardless of our hypothesis being true, we have covered numerous points of evidence and argued in depth that the senescent glial cell burden contributes to multiple pathological aspects of AD in mouse models and patients: spreading protein aggregates of amyloid beta (Aβ) and/or hyperphosphorylated tau (hp‐tau), demyelination, synaptic loss, and spatial memory performance. ³⁴

While multiple glial types likely become senescent via AD pathology and contribute to further AD progression, both mouse and human microglia have been particularly shown to induce senescence via hp‐tau uptake, whether by endocytosing hp‐tau aggregates secreted from other microglia or by phagocytosing neurons containing hp‐tau aggregates. ⁸³ Both mouse and human senescent microglia have been shown to upregulate ferritin, indicating iron accumulation. ¹⁹⁰ Although also found specifically in neuritic plaques, ¹⁹² iron accumulation in microglia can serve as a senescence hallmark; ¹³⁴ , ¹³⁵ iron accumulation additionally well correlates with AD Braak staging, and comprises a strong independent predictive factor for cognitive decline. ¹⁹³ , ¹⁹⁴

Although the nuances and likely heterogeneity of senescent microglia are addressed further in our previous review, senescent microglia take on a dystrophic spectrum indicated by a progressive reduction in morphological complexity. ³⁴ , ¹⁹⁰ More advanced microglial dystrophy in humans also involves spheroidal swellings and possible cytoplasmic degeneration, without signs of apoptosis. ³³ , ¹⁹¹ Dystrophic microglia have been found to precede and associate with neurofibrillary tangle (NFT) pathology, ³³ and we also argued that ferritin‐enriched, senescent microglia likely (1) display varying degrees of hypertrophy with their dystrophic morphologies, with (2) extremely hypertrophic senescent and dystrophic microglia having enriched association with neuritic plaques. ³⁴ Combining that neuritic plaques predominate hp‐tau spreading over NFT abundance and other forms of neuron‐based hp‐tau aggregation, ³⁶ and that microglia facilitate amyloid plaque formation, ³⁸ it follows that AD likely involves accelerated neurofibrillary pathology spreading via senescent microglia creating neuritic plaques (containing hp‐tau). ³⁴

The requirements for accumulating microglial senescence and hp‐tau spreading are not ordinary features of cognitively healthy individuals. Instead, microglia and initial hp‐tau pathology likely become primed for interaction in preclinical AD due to possible aging effects, including declining homeostatic glial states, slow increases to ongoing inflammatory signaling and oxidative stress, recurrent AD pathology creation and buildup, and senescence induction in other glial types. ³⁴ , ⁸³ , ⁹⁸ , ¹³⁹ , ¹⁷⁵ , ¹⁸³ , ¹⁸⁴ , ¹⁹⁰ Nonetheless, the spatiotemporal progression of glial senescence in AD is predicted to matter especially when microglial senescence and accompanying hp‐tau spread in regions relevant to AD Braak staging (Box 1). As argued here, the unique combination of causes and factors sustaining glial senescence in other brain regions may cause other clinical symptoms; concurrent senescence accumulation in multiple brain regions may emergently account for co‐morbid pathology and symptoms in patients displaying multiple neurodegenerative diseases, including dementia with Lewy bodies.

Here, we will review evidence that dystrophic microglia are senescent. In primary mouse microglial cell cultures exposed to ferric citrate, microglia took on a dystrophic morphology (reduced morphology complexity) with ferritin upregulation. ⁷⁹ In 5xFAD mice, microglia with senescence markers (lowered Lamin B1; increased HMGB1, γH2A.X, and SA‐β‐gal) were found to display dystrophic morphologies via decreased morphological complexity. ⁷² Human microglia within gray and white matter in the neocortex displayed dystrophic morphologies (indicated by enriched ferritin light chain staining, or including iron accumulation as another senescence marker, ¹³⁴ , ¹³⁵ decreased ramification, and appearance with spheroidal swellings) also displayed increased senescence‐associated yH2A.X. ¹⁹⁵

Jin et al. have demonstrated increased type I IFN signaling and synaptic pruning by iPSC human and mouse senescent and dystrophic microglia (dystrophy being defined via displaying both ferritin enrichment and decreased morphological complexity) displaying enrichment in senescence‐associated genes, which mediated neurodegeneration particularly in an immunodeficient mouse model injected with hp‐tau aggregates from patient brain tissue. ¹⁹⁰ While unconfirmed in PD without dementia, increased synaptic pruning and subsequent neurodegeneration may be another form of damage elicited by senescent microglia in LBD. It is also currently unknown how senescent microglia may arise in LBD, but one potential route may include “phagoptosing” or prematurely phagocytosing stressed neurons and neurites, including a subset of senescent neurons/neurites containing both α‐syn and hp‐tau aggregates (Figure 3). ²⁶ , ⁴¹ , ⁸³ , ¹¹⁵ Aside from human and mouse microglia displaying senescence markers (including ferritin [indicating iron] enrichment, several transcriptomic genes for type I IFN signaling, and B2M, ZFP36L1, and XAF1 upregulation) in vivo after exposure to hp‐tau from patient brain, ¹⁹⁰ overexpressing P301S tauopathy mouse models and primary mouse microglial cells can present senescence markers (including lower Lamin B1 expression, and increased p16^INK4a and SA‐β‐gal levels). ⁸³ , ⁸⁴ , ¹⁹⁶ Primary mouse microglial cells were shown to become senescent after phagoptosing neurons containing hp‐tau aggregates, whereupon senescent microglia‐derived conditioned medium separately induced senescence in control microglial cells. ⁸³ As multiple environmental insults have been discussed above to also induce hp‐tau pathology, ¹¹¹ , ¹¹⁵ , ¹²³ , ¹⁵⁵ , ¹⁵⁶ , ¹⁵⁷ , ¹⁵⁸ it likely follows that microglial senescence can occur in LBD and contribute indirectly to α‐syn aggregation (Figure 3). Senescent microglia may additionally contribute to the increased cGAS‐cGAMP‐STING signaling and neurodegeneration observed in PD, ¹⁸⁷ via upregulation of type I IFN responses innately connected to cGAS‐cGAMP‐STING signaling. ³⁴ , ⁶⁹ Finally, as α‐syn fibrils were shown to upregulate senescence markers in primary microglial cells (increased p21, lowered Lamin B1), ¹⁴⁰ hp‐tau may speculatively serve as a strong driving force in inducing microglial senescence versus α‐syn and other protein aggregates.

Although more studies are required, senescent glial cells have been documented in samples from patients with PD and DLB. Increased IL6, IL8, and p16^INK4a RNA levels were confirmed in the SNpc of post mortem PD patient tissue with decreased Lamin B1 immunostaining in GFAP+ positive cells, ¹¹² indicating senescent astrocyte enrichment. Dystrophic microglia were similarly observed in the striatum of patients with PD, ¹⁹⁷ and in multiple regions within patients with DLB. ¹⁹⁸ , ¹⁹⁹ , ²⁰⁰ In the hippocampus, dystrophic microglia were quantified as overrepresented in post mortem samples from patients with DLB versus individuals with intact cognition. ²⁰⁰ However, dystrophic microglia were not as overrepresented in the superior frontal gyrus of patients with DLB and diagnosed AD Braak stage II to IV versus cognitively healthy controls; as the frontal lobe shows less hp‐tau burden until AD Braak stage V, ³¹ this is likely due to significant hp‐tau burden and accompanying, predicted senescent dystrophic microglia burden not having reached the superior frontal gyrus (via increased synaptic pruning, accelerating further hp‐tau spread, and creating neuritic plaques containing hp‐tau). ³⁴ , ⁸³ , ¹⁹⁰ , ¹⁹⁶ We predict here that senescent microglia burden and contributions to LBD progression may less involve progressing toward new regions, instead likely more toward initiating neuronal senescence, decreasing protein aggregate clearance (including α‐syn), and promoting LB formation in previously affected regions. ⁷⁹

Notably, post mortem examinations have revealed that patients with DLB do not all display significant amounts of CSF hp‐tau or progress past hp‐tau burden indicating AD Braak stage III (correlating to mid‐late stage progression of AD ³¹ , ³⁴ ). ⁴⁶ , ⁵⁴ , ⁵⁷ , ⁵⁸ Additionally, manganese (i.e., a heavy metal) exposure does not induce LB formation but has been well established to trigger cognitive impairment, parkinsonian symptoms, and both Aβ and hp‐tau aggregation. ²⁰¹ , ²⁰² , ²⁰³ Particularly, dystrophic microglia were confirmed to be found within the substantia nigra of macaques exposed chronically to manganese. ²⁰⁴ As these macaques demonstrate parkinsonian symptoms, senescent microglia may also contribute to parkinsonian and other motor deficits seen later in patients with end‐stage dementia. ⁴² , ⁴³

Finally, we have also not covered MSA, which involves α‐syn pathology and spread primarily involving oligodendrocytes and neurons. ²⁰ , ²¹ Speculatively, even though hp‐tau aggregation does not play a main role in MSA, both manganese exposure and MSA can lead to increased amounts of iron‐rich microglia and cognitive decline. ²² , ²⁰⁴ , ²⁰⁵ , ²⁰⁶ Considering this and the previous argument for dementia being advanced by threshold glial (and especially microglial) senescence in AD Braak staging areas, ³⁴ and senescent microglia being iron rich, ³⁴ , ⁷⁹ , ¹⁹⁰ senescent (micro)glia accumulation may even sufficiently explain and drive cognitive decline independently of hp‐tau burden. This also includes extending to, and accounting for, cases of cognitive impairment in DLB without significantly enriched AD‐associated pathology burden; ⁴⁶ , ⁵⁴ , ⁵⁷ glia, including astrocytes and microglia, can undergo senescence even without overwhelming exposure to such Aβ and hp‐tau pathologies. ¹⁸⁶ , ¹⁸⁸ , ²⁰⁷ Although requiring further investigation, our hypothesis accounts for cognitive impairment in such non‐AD associated DLB cases; such DLB variants should result via a buildup of glial and neuronal senescence enriched in AD Braak staging areas.

1.4. Local environmental insults and neuronal senescence are likely primary contributors to PD spread and heterogeneity

Environmental insults and senescence spread are hypothesized to explain and predict late‐onset and environmental PD progression (i.e., not originating from the brain‐first SOC hypothesis). ¹ , ¹⁴ To resolve differences between the “body‐first” SOC and USSLBD, ⁹ , ¹⁰ , ¹⁴ we propose here that initial α‐syn pathology and LB spread originate from the olfactory bulb (USSLBD Stage I), enteric nervous system (SOC), or both regions (Figure 4). As senescence spread occurs via paracrine mechanisms, senescence spread would necessitate local connections between cells; ⁷⁰ here, senescence spread involving α‐syn pathology putatively involves neuronal connectome and/or paracrine spread from senescent glia. ² , ⁸³ Therefore, α‐syn pathology in the gut is predicted to occur independently or simultaneously with α‐syn pathology in the olfactory bulb; environmental insults may also (co‐)occur independently through nasal and/or gut and vagus nerve exposure into the body. ¹⁰⁴ , ¹⁰⁷ , ¹¹⁷ , ²⁰⁸ We predict that initial LBD development occurs due to chronic environmental insult exposure; ¹⁰⁷ , ¹¹⁴ , ¹¹⁷ minor challenges from environmental insults over time likely result in decreased cell death and senescence that can be cleared out by healthy glial cells, but these systems falter with aging, ³⁴ individual physiology and genetic risk factors, and/or repeated environmental insult exposure (Figure 4). Senescent glia, whether induced directly by environmental insults, aging, and/or via co‐occurring neurodegenerative disease, are also suspected to help initiate neuronal senescence and permit downstream α‐syn pathology spread. ⁷³ Finally, the gut–brain route may involve macrophages that also become senescent and permit α‐syn pathology spread via endolysosomal stress. ²⁰⁹

α‐Syn pathology propagation into unaffected regions putatively occurs mainly via neuronal senescence, ⁷³ , ¹²⁸ as neurons connect directly to each other and selectively vulnerable neurons in LBD often have long axons, ² new regions through a synaptic connectome—simultaneously passing α‐syn fibrils and inducing both ILBD and senescence into distal, selectively vulnerable neurons. ² , ¹³² Furthermore, the olfactory bulb is located closer to limbic regions and the enteric nervous system closer to brainstem regions. As senescence requires local spread, ³⁴ , ⁶⁹ , ⁷⁰ environmental PD is proposed to involve “olfactory” and “gut–brain” routes: the olfactory route begins in the olfactory lobe and spreads into limbic‐predominant regions before brainstem involvement (USSLBD IIb), whereas the gut–brain route involves progressing into brainstem regions via spreading through the vagus nerve (USSLBD IIa; ⁹ , ¹⁰ Figure 4).

Per the SOC model, the body‐first and brain‐first subtypes accumulate differing extents of α‐syn prior to clinical symptom detection; the body‐first subtype likely contains greater amounts of hemisphere‐symmetrical and global α‐syn burden, relative to the brain‐first subtype. ¹ , ¹⁴ Although the olfactory route is anatomically closer to the SNpc than the gut–brain axis would be, increased α‐syn burden should accompany increased proportions of senescent cells; thus, while we overview that neuronal senescence primarily facilitates α‐syn pathology into new regions, we also predict that there must be sufficient senescence buildup before a region becomes impaired to the extent of contributing to significant, clinical impairment. Here, we can take that increased, but subclinical burden of senescent cells and α‐syn pathology across multiple regions, as in the case of the brain‐first SOC subtype, should then facilitate much faster decline and spread between regions when a first region (olfactory bulb or vagus nerve) reaches clinical threshold (i.e., sufficient burden of both senescent cells and α‐syn pathology). This also corresponds with senescence occurring in development, but with senescent cell numbers in development not persisting chronically (i.e., said senescent cells are cleared out by immune cells), in contrast to disease conditions (i.e., senescent cells persisting, building up, and contributing to disease pathologies). ⁷⁰ , ¹⁹⁶ , ²¹⁰ , ²¹¹

Irrespective of initial route, the olfactory and/or gut–brain routes are then predicted to spread into previously unaffected brainstem or limbic regions; this should correspond and agree with SNpc neurodegeneration and parkinsonian symptoms by USSLBD III. ⁹ , ¹⁰ Here, subsequent α‐syn pathology proposedly spreads into the neocortex (USSLBD IV), indicating more end‐stage PD and increased dementia risk (Figure 4). Environmental PD progression should conform with our predictions overall, and yet display local heterogeneity of spread depending on an individual's physiology, genetics, and lifestyle factors.

1.5. Cell type–specific senescence spread could novelly account and differentiate progression rates and dementia prognosis in LBD and AD

Due to positive feedback loops of environmental insults, α‐syn pathology, and AD pathology burden, co‐occurring AD pathology buildup is predicted to happen over LBD progression. ⁴⁸ , ⁹⁹ , ¹⁰⁸ , ¹¹¹ , ¹¹⁵ , ¹⁵³ , ¹⁵⁵ , ¹⁵⁶ , ¹⁵⁷ , ¹⁶² If this is true, what differentiates AD from PD? Here, senescence spread is proposed to explain both overlap and distinguishing factors between PD and AD in the LBD spectrum. Particularly, we predict that the rate of an individual's parkinsonian symptoms and/or dementia prognosis arises from the (1) senescence accumulation rates of multiple cell types within and between disease regions, and (2) the identity of the spatial regions presenting critical senescence burden.

Regarding PD progression, environmental insults and/or α‐syn pathology buildup have already been argued to involve (1) neuron‐autonomous senescence spread caused by α‐syn release between neurons and direct connectomes in (2) (subsequent) regions indicated in Figure 4. ² , ¹³² We have also already discussed that environmental insults and α‐syn pathology buildup induce glial senescence, with stated glial senescence decreasing neuronal viability and initiating neuronal senescence. ⁷⁹ , ¹¹² , ¹⁸⁶ As the extent of glial senescence spreading α‐syn pathology without senescent neurons has not been studied, and neuronal propagation of α‐syn pathology via LB appearance and connectomes is supported by the literature, ¹ , ² , ⁹ , ¹³ neuronal senescence remains as a predicted, primary facilitator of α‐syn pathology into new regions.

This contrasts with AD, in which initial hp‐tau pathology and NFTs from (senescent) neurons are necessary, but facilitate slower hp‐tau seeding and spread less toward AD progression versus hp‐tau derived from neuritic plaques; ³⁶ neuritic plaques are built by microglia, which we argued are senescent microglia due to their phagocytosis and secretion of hp‐tau ³⁴ , ⁸³ , ⁸⁴ (Box 2). Dystrophic microglia and neuritic plaques also appear earlier in abundance compared to significant NFT burden in AD, ³³ , ³⁶ can cause DLB without significant NFT burden in the cortex, ²¹² and hold more importance in determining clinical presentation than LB burden for patients with DLB. ²⁹ Additionally, while environmental insults were discussed with regard to inducing α‐syn pathology, different toxins can induce Aβ and hp‐tau aggregation; ¹¹¹ , ¹⁵⁵ , ¹⁵⁶ , ²¹³ , ²¹⁴ , ²¹⁵ , ²¹⁶ , ²¹⁷ notably, precise biochemical mechanisms inducing such aggregation are not clearly understood across every single environmental insult. Nonetheless, how PD versus AD spreads can now be further comprehended and differentiated via environmental insult exposure.

To explain this, we will cover example cases of “pure,” or non‐overlapping variants of PD and AD progression; without accounting for real‐life overlap between these diseases, we will consider a patient with pure PD and no overlapping AD pathology, and an alternative case of a patient with AD without any α‐syn pathology. In PD, both entry routes are predicted to involve the earliest accumulation of significant LB at peripheral entry sites (i.e., via the olfactory bulb or vagus nerve); ¹ , ⁹ , ¹⁰ , ¹⁴ in a pure PD trajectory without any comorbid diseases, we would predict that senescence buildup occurs first in these peripheral sites. In contrast, early progression in pure AD trajectories should not involve similar extents of environmental insult exposure and subsequent damage to peripheral entry routes (i.e., damage being a greater number of neurons and glia undergoing either cell death or senescence, with the resulting population of neurons and glia having undergone an increased total number of dead and senescent cells). “Pure” AD progression may instead arise out of genetic factors (e.g., apolipoprotein E ε4 genotype carriers) or cerebrovascular pathology (e.g., stroke damaging the transentorhinal cortex).

While environmental insults likely induce systemic inflammatory signaling damaging regions involved in early AD Braak staging (e.g., transentorhinal cortex), ²¹⁸ , ²¹⁹ the temporal lobe is a more anatomically difficult area for environmental insults to reach, versus the anterior olfactory nucleus or vagus nerve. As environmental insults increase both PD and AD risk, ¹⁰⁷ , ¹¹² , ¹²¹ , ²¹⁹ likely via inducing the aforementioned systemic inflammatory signalling, ²¹⁸ , ²¹⁹ sufficient environmental insult exposure should also produce mixed disease with both α‐syn and AD pathologies. ¹¹³ Here, α‐syn aggregation in environmental insult‐initiated, mixed disease progression will likely occur first before overt symptomatic detection, in due part via these environmental insults; later increased α‐syn burden seeding and sufficient senescence burden should then fall in line with predictions made via the SOC model. ¹ , ¹⁴ Finally, even if such mixed disease progression does not determine a patient as demonstrating DLB, any AD presentation involving exposure to environmental insults should not be “pure”; such AD variants are predicted to exist with some level of increased α‐syn aggregation in the olfactory and/or vagus nerves.

Here, we can build on of our previous arguments and literature discussion that microglial senescence burden in AD Braak staging regions acts as a primary driver of dementia progression; ³⁴ dementia progression refers to the spread of neurodegenerative pathologies and accompanying symptomatic impairment, as outlined via in Braak staging regions used to classify AD diagnosis. ³¹ Regardless of α‐syn pathology, we can logically follow that (1) glial (requiring microglial) senescence is a greater predictor of AD Braak staging regions versus (2) neuronal senescence within enteric nervous system and USSLBD regions are predicted to differentiate relative (1) AD and/or (2) PD progression (Figure 1). ⁴⁴ PDD and DLB are considered here as unique combinations of senescence accumulation and spread rates involving hp‐tau versus α‐syn pathology, where there is increased glial senescence (requiring threshold senescent microglia accumulation) in AD Braak staging regions; ³⁴ PD involves appearance and greater first spread of α‐syn pathology via neuronal senescence, compared to DLB involving accelerated involvement and spread via (micro)glial senescence accumulation. ²⁹ , ⁴⁵

Our predictions also account for motor symptoms that become more prevalent in late‐stage AD. ⁴² , ⁴³ Explicitly, although we predict that neuronal senescence and α‐syn pathology occur prior to significant glial senescence, the numbers of such neuronal senescence and α‐syn burden may require further buildup before progressing into threshold clinical detection; this corresponds to α‐syn buildup occurring much more in the SOC model's body‐first PD subtype, before clinical detection. ¹ , ¹⁴ Thus, in both cases of DLB or body‐type first PD, increased amounts of pre‐existing senescence burden (where such burden may already be detectable symptomatically, but does not yet satisfy clinical diagnosis) are predicted to induce faster clinical progression into end‐stage LBD (Figure 1).

Finally, DLB is also accompanied by early indications distinctly separating it from AD, including enriched visual hallucinations and depressive symptoms. ²²⁰ , ²²¹ , ²²² Briefly, these neuropsychiatric symptoms likely result via impaired involvement from multiple regions; hallucinations have been reported to be associated with reduced gray matter throughout the cerebral cortex and cingulate gyrus, ²²³ whereas depression involves both cortical and especially limbic lobe dysfunction. ²²⁴ Finally, although not yet empirically investigated, we would predict that senescence buildup would be increased within regions responsible for increased hallucinations and depressive symptoms. For instance, patients with DLB with hallucinations would be predicted to have increased senescent glia and neurons in the inferior parietal lobule, cuneus, and lateral occipital cortex versus patients with AD; all these regions should be further impaired in patients reporting hallucinations, ²²³ and should reflect increased senescent burden in these corresponding areas. As different stressors (in α‐syn, Aβ, and hp‐tau pathologies) can all induce senescence, ¹²⁸ , ¹⁴⁴ , ¹⁴⁵ , ¹⁸⁸ it is likely these pathologies should also be increased and locally accompany the predicted, increased senescence burden in regions exclusive to DLB (e.g., increased neuritic plaques, α‐syn aggregates, and senescent cells in the cuneus). Notably, this should also be further investigated empirically.

2. CONCLUSION

We posit that the LBD spectrum comprises diseases which are advanced by environmental insults and senescence spread (Figure 4). LB pathology and α‐syn pathology spread is predicted to rely primarily on the spreading of senescence via the connectomes of selectively vulnerable neurons, in brain areas illustrated in Figure 4; ¹ , ² , ¹⁰ overlapping dementia progression and symptoms are primarily driven by senescent glial burden (particularly requiring threshold microglial senescence) in AD Braak staging brain regions. ³¹ , ³⁴ According to our hypothesis, the spatiotemporal progression of senescence accumulation and spread would dictate an overt patient diagnosis of (overlapping) PD, PDD, AD, or DLB (Figure 1).

Other risk factors contributing to PD are not fully explored here, but deserve further, separate investigations. These reductively include effects of environmental insults, α‐syn aggregates, aging, and senescence on the enteric nervous system, peripheral immune system, and infiltrating immune cells; however, all other factors relevant to PD progression are presumed to play critical roles toward shaping the convergent damaging effects of environmental insults, senescence spread, and co‐occurring α‐syn and AD pathology observed in LBD (Figures 1, 2, 3). While LRRK2, SNCA, and GBA have been briefly explored in their contributions to neuronal senescence and LBD progression, ¹³⁰ , ¹⁴⁷ , ¹⁸⁶ other genetic risk factors, including PARK7, PINK1, and PARK2, may converge to permit increased oxidative stress. ¹² This increased oxidative stress likely raises the selective vulnerability of neurons involved in LBD, ² ultimately increasing the risk of LBD initiation.

Given the complexity of these overlapping diseases and multiple senescent cell types being affected concurrently, using senolytics to selectively kill senescent cells appears attractive. ³⁴ , ²²⁵ , ²²⁶ While senolytics do not likely stop the base incident rate of environmental insults and initial α‐syn and/or AD pathology from forming, senescent cells have already been discussed earlier to exert disproportionately anti‐supportive and damaging effects on other non‐senescent, homeostatic populations. As senescent cells induce senescence in non‐senescent cells, ³⁴ , ⁶⁹ , ⁷⁰ senolytic treatment should critically decrease the rate of acceleration leading into LBD and dementia progression. Senolytic treatment is also expected to increase the quality of life in patients along the LBD and AD spectrum, as patient brain samples often display co‐morbid α‐syn and AD pathology. ²⁶ , ²⁸ , ²⁹ , ⁴⁷ In addition, promoting autophagy or minimizing autophagic dysfunction may also help to minimize senescent cell buildup. ⁸⁵

Finally, senolytics are recommended to be combined with add‐on treatments, whether holistic and/or pharmacological, to further target pathologies selective or co‐occurring across both PD and AD. ³⁴ Exploiting strategies of senescent cells, such as possessing decreased lysosomal function, has led to developing chiral gold nanoparticles accumulating in senescent cells; in A53T mice, exposing these nanoparticles to light ablated an enriched senescent microglial population and demonstrated simultaneous motor and memory improvement. ¹⁸⁹ Thus, targeting senescent burden may allow for optimally targeting multiple aging, neurodegenerative diseases impairing quality of life. Other treatments and lifestyle modifications may act as preventative measures for treating LBD, ³ , ²²⁷ , ²²⁸ via minimizing senescent cell induction driving LBD progression.

AUTHOR CONTRIBUTIONS

Victor Z. Lau led article conceptualization, manuscript writing, revising, and figure creation. Ifeoluwa O. Awogbindin, Shawn N. Whitehead, Dan Frenkel, and Marie‐Ève Tremblay contributed to revising and improving the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

Consent was not necessary.

Supporting information

Supporting Information

Lau VZ, Awogbindin IO, Frenkel D, Whitehead SN, Tremblay M‐È. A hypothesis explaining Alzheimer's disease, Parkinson's disease, and dementia with Lewy bodies overlap. Alzheimer's Dement. 2025;21:e70363. 10.1002/alz.70363